Sentimetrix: A Robust Sentiment Analysis Framework Using Ensemble Learning Transformer with Lime Interpretability
Keywords:
Sentimetrix, Hybrid Transformers, Ensemble Learning, BERTweet, VADER, TextBlob.Abstract
This study introduces Sentimetrix, a sophisticated sentiment analysis system that aims to extract subtle opinions and feelings from textual data. Sentimetrix uses not only the latest state-of-the-art techniques in sentiment classification, including hybrid transformers
(BERT, BERTweet, RoBERTa) and RNN models, but also ensemble learning to achieve high accuracy and interpretability.
References
Sorato, D., Lundsteen, M., Ventura, C. C., & Zavala-Rojas, D. (2024). Using word embeddings for immigrant and refugee stereotype quantification in a diachronic and multilingual setting. Journal of Computational Social Science, 1-53.J. Clerk Maxwell, A Treatise on Electricity and Magnetism, 3rd
ed., vol. 2. Oxford: Clarendon, 1892, pp.68–73.
Krols, T. L., Mortensen, M., & Oldenburg, N. (2023). Profiling Irony & Stereotype: Exploring Sentiment, Topic, and Lexical Features. arXiv preprint arXiv:2311.04885.
Toshevska, M., Kalajdziski, S., &Gievska, S. (2023, September). Graph Neural Networks for Antisocial Behavior Detection on Twitter. In International Conference on ICT Innovations (pp. 222-236).
Cham: Springer Nature Switzerland.
